Background Obesity-related SDB requiring domiciliary non-invasive ventilation (NIV) can present as (1) eucapnic obstructive sleep apnoea (OSA) (2) hypercapnic OSA (P_aCO₂ >6 kPa) (3) hypercapnic OSA with obesity hypoventilation syndrome (OHS) and (4) lone OHS. We have adopted the term obesity-related respiratory failure (ORRF) to group these differing conditions. The aim of this study was to determine the clinical features prevalent in each of these conditions in order to guide respiratory management.

Method Data from patients initiated on domiciliary NIV at a tertiary referral unit, between August 2005 and December 2009, were obtained from a purpose-designed discharge summary database. Patients were categorised into four groups, as described above. Comparative analyses were performed between (1) eucapnic OSA and the hypercapnic groups (2) hypercapnic OSA and a group combining the OSA with OHS and the lone OHS groups (OSA & OHS and lone OHS) and (3) eucapnic and hypercapnic OSA. Logistic regression analysis was performed to determine factors associated with hypercapnia.

Results 163 patients were included in the analyses. Group mean (SD) age 54.3±14.2 years, weight 134.4±33.1 kg, body mass index (BMI) 48.4±2.5 kg/m² and Epworth sleepiness (ESS) score 14.8±5.8. Results are shown in Abstract S67 Table 1. The hypercapnic groups demonstrated a higher prevalence of diabetes. In addition, hypercapnic patients were overall, compared with eucapnic patients, more hypoxic with greater lung restriction, despite a non-significant increase in BMI in the hypercapnic OSA group. Compared with the hypercapnic OSA group, the combined OSA and OHS and lone OHS group had a higher BMI, ESS and greater hypercapnia. However, logistic regression analysis failed to demonstrate any factors that predicted hypercapnia.

Conclusion Demographic, anthropometric, spirometric and clinical features allow the different ORRF conditions to be distinguished. BMI, daytime symptoms and degree of chronic respiratory failure distinguished between hypercapnic OSA from obesity-related hypoventilation. Although we hypothesise that ORRF is a disease spectrum, from eucapnic OSA progressing to hypercapnic OSA to OSA with OHS, we were unable to identify factors that predicted hypercapnia. From these data, we propose that more detailed physiological assessment, including neural respiratory drive and pulmonary mechanics, is required.